GAO Says States Are Underpaying Unemployment Claims


The Government Accountability Office says the jobless benefits program consistently underpaid some workers.

Flawed Estimates 

The GAO claims the Labor Department Publishes ‘Flawed Estimates’ of Weekly Jobless Claims.

Weekly news releases issued by the Department of Labor (DOL) do not provide an accurate estimate of the total number of individuals actually claiming unemployment insurance (UI) because they have potentially both over-estimated and underestimated the total number of individuals actually claiming unemployment insurance. That is because DOL presents state-reported data on the total count of weeks claimed, nationwide, as the number of people claiming benefits. DOL has traditionally used the total count of weeks claimed to estimate the number of individuals claiming benefits because the two numbers were a good approximation of each other. However, due to state backlogs in processing claims and other data issues, these traditional estimates are not appropriate in the context of the pandemic. For example, state backlogs in processing claims led to individuals submitting claims for multiple weeks of retroactive benefits during single reporting periods. So, by using claims counts to represent the number of people, many individuals are counted more than once in DOL’s estimate.

In addition, DOL has also inaccurately characterized changes in claims numbers from week to week, in large part because of inconsistency in the group of states reporting Pandemic Unemployment Assistance (PUA) data each period. For example, in its July 23, 2020, publication, DOL reported that the number of claims in all programs for continuing unemployment during the week ending July 4 decreased by about 200,000 from the previous week. However, Arizona did not report PUA data for that week, after reporting almost 2.3 million claims the previous week. Had Arizona submitted data, DOL likely would have reported a significant increase in claims from the prior week instead of a decrease. To understand the role UI benefits are playing in the economy during the pandemic, reliable data are needed on both the number of new claimants each week and the number of claimants who continue to rely on UI benefits. DOL should revise its weekly news releases to clarify that in the current unemployment environment, the numbers it reports for weeks of unemployment claimed do not accurately estimate the number of unique individuals claiming benefits. DOL agreed.

 Flawed Estimates and Flawed Payouts 

The Wall Street Journal comments on Flawed Estimates and Flawed Payouts.

The Labor Department’s weekly reports on jobless claims have published “flawed estimates of the number of individuals receiving benefits each week throughout the pandemic,” the Government Accountability Office said in a periodic report.  

In addition, a program created by Congress to provide jobless benefits to workers who are normally not eligible for them has underpaid recipients in most states. As a result, the average weekly payout under what is called the Pandemic Unemployment Assistance program, which is available to gig-economy workers and the self-employed, is below the poverty line in 70% of states that reported data.

The majority of states have been paying PUA claimants the minimum allowable benefit instead of the amount they are eligible for based on prior earnings,” the GAO said. While states are obligated to pay out the full amount that is owed “with the greatest promptness that is administratively feasible,” Labor Department officials told the GAO they didn’t know how many states had begun working to do so.

“Without an accurate accounting of the number of individuals who are relying on [unemployment insurance] and PUA benefits in as close to real-time as possible, policy makers may be challenged to respond to the crisis at hand,” the GAO said.

Unemployment Trends: How Many Layoffs are Temporary?

Unemployment Level in Thousands 2020-10

If you seek undercounts, take a look at how the BLS classifies the unemployed.

For discussion, please see Unemployment Trends: How Many Layoffs are Temporary?

Number of People Unemployed for 27 Weeks or More

Number Unemployed 27 weeks or Longer October

Of those unemployed, there is a Dramatic Surge in Percentage of Long-Term Unemployment.

No Thanksgiving Cheer

There is No Thanksgiving Cheer for the Unemployed.

All Continued Claims for  2020-11-25 Report

Click on the above link for the current breakdown of Initial Claims, Continued Claims, PUA Claims, and PEUC Claims.

Benefits Expire December 26

PUAC and PEUC expires on December 26.

About 3.5 million people already exhausted PEUC.

A fiscal cliff looms for the unemployed.


Comments (5)
No. 1-5



Thanks for the data. Gonna be a rough 2021


Just goes to show how bright most of these employees are who are managing these programs, one should always be looking for "win-win" transactions where at all possible.

There is no more fiscal prudence or anything of the sort, it's over and if you think it exists you are screwed, one should be trying to get each and every person their maximum payout as that money will most likely be spent and circulated in the economy, this is the same logic as tax cuts, the money will be invested/spent/etc back into the economy mantra. Does it work, who knows?

The printing presses are just getting started, they haven't even warmed up yet, it's a tricky balance as a loss of confidence event could bring it all down, so tremendous amounts of money must be funneled to the everyday person if the show is to go on.

Lance Manly
Lance Manly

Yeah, very similar to the issues they had with seasonality at the beginning of the crisis, the normal methodology did not apply. Would have been better if they could have pivoted more agily, but also then you get naysayers that claim fraud because you are changing the methodology. I would like to see the GAO audit the BLS, that would be fun. I doubt very many of their models performed well under these circumstances.

Global Economics